Name

unicode_u_ucs4_native, unicode_u_ucs2_native, unicode_convert_init, unicode_convert, unicode_convert_deinit, unicode_convert_tocbuf_init, unicode_convert_tou_init, unicode_convert_fromu_init, unicode_convert_uc, unicode_convert_tocbuf_toutf8_init, unicode_convert_tocbuf_fromutf8_init, unicode_convert_toutf8, unicode_convert_fromutf8, unicode_convert_tobuf, unicode_convert_tou_tobuf, unicode_convert_fromu_tobuf — unicode character set conversion

Synopsis

#include <courier-unicode.h>

            extern const char unicode_u_ucs4_native[];

            extern const char unicode_u_ucs2_native[];
unicode_convert_handle_t unicode_convert_init( const char *src_chset,
  const char *dst_chset,
  void *cb_arg);
 
int unicode_convert( unicode_convert_handle_t handle,
  const char *text,
  size_t cnt);
 
int unicode_convert_deinit( unicode_convert_handle_t handle,
  int *errptr);
 
unicode_convert_handle_t unicode_convert_tocbuf_init( const char *src_chset,
  const char *dst_chset,
  char **cbufptr_ret,
  size_t *cbufsize_ret,
  int nullterminate);
 
unicode_convert_handle_t unicode_convert_tocbuf_toutf8_init( const char *src_chset,
  char **cbufptr_ret,
  size_t *cbufsize_ret,
  int nullterminate);
 
unicode_convert_handle_t unicode_convert_tocbuf_fromutf8_init( const char *dst_chset,
  char **cbufptr_ret,
  size_t *cbufsize_ret,
  int nullterminate);
 
unicode_convert_handle_t unicode_convert_tou_init( const char *src_chset,
  char32_t **ucptr_ret,
  size_t *ucsize_ret,
  int nullterminate);
 
unicode_convert_handle_t unicode_convert_fromu_init( const char *dst_chset,
  char **cbufptr_ret,
  size_t *cbufsize_ret,
  int nullterminate);
 
int unicode_convert_uc( unicode_convert_handle_t handle,
  const char32_t *text,
  size_t cnt);
 
char *unicode_convert_toutf8( const char *text,
  const char *charset,
  int *error);
 
char *unicode_convert_fromutf8( const char *text,
  const char *charset,
  int *error);
 
char *unicode_convert_tobuf( const char *text,
  const char *charset,
  const char *dstcharset,
  int *error);
 
int unicode_convert_toubuf( const char *text,
  size_t text_l,
  const char *charset,
  char32_t **uc,
  size_t *ucsize,
  int *error);
 
int unicode_convert_fromu_tobuf( const char32_t *utext,
  size_t utext_l,
  const char *charset,
  char **c,
  size_t *csize,
  int *error);
 

DESCRIPTION

unicode_u_ucs4_native[] contains the string UCS-4BE or UCS-4LE, matching the native char32_t endianness.

unicode_u_ucs2_native[] contains the string UCS-2BE or UCS-2LE, matching the native char32_t endianness.

unicode_convert_init(), unicode_convert(), and unicode_convert_deinit() are an adaption of th iconv(3) API that uses the same calling convention as the other algorithms in this unicode library, with some value-added features. These functions use iconv(3) to effect the actual character set conversion.

unicode_convert_init() returns a non-NULL handle for the requested conversion, or NULL if the requested conversion is not available. unicode_convert_init() takes a pointer to the output function that receives receives converted character text. The output function receives a pointer to the converted character text, and the number of characters in the converted text. The output function gets repeatedly called, until it receives the entire converted text.

The character text to convert gets passed, repeatedly, to unicode_convert(). Each call to unicode_convert() results in the output function getting invoked, zero or more times, with each successive part of the converted text. Finally, unicode_convert_deinit() stops the conversion and deallocates the conversion handle.

It's possible that a call to unicode_convert_deinit() results in some additional calls to the output function, passing the remaining, final parts, of the converted text, before unicode_convert_deinit() deallocates the handle, and returns.

The output function should return 0 normally. A non-0 return indicates n error condition. unicode_convert_deinit() returns non-zero if any previous invocation of the output function returned non-zero (this includes any invocations of the output function resulting from this call, or prior unicode_convert() calls), or 0 if all invocations of the output function returned 0.

If the errptr is not NULL, *errptr gets set to non-zero if there were any conversion errors -- if there was any text that could not be converted to the destination character text.

unicode_convert() also returns non-zero if it calls the output function and it returns non-zero, however the conversion handle remains allocated, so unicode_convert_deinit() must still be called, to clean that up.

Collecting converted text into a buffer

Call unicode_convert_tocbuf_init() instead of unicode_convert_init(), then call unicode_convert() and unicode_convert_deinit() normally. The parameters to unicode_convert_init() specify the source and the destination character sets. unicode_convert_tocbuf_toutf8_init() is just an alias that specifies UTF-8 as the destination character set. unicode_convert_tocbuf_fromutf8_init() is just an alias that specifies UTF-8 as the source character st.

These functions supply an output function that collects the converted text into a malloc()ed buffer. If unicode_convert_deinit() returns 0, *cbufptr_ret gets initialized to a malloc()ed buffer, and the number of converted characters, the size of the malloc()ed buffer, get placed into *cbufsize_ret.

Note

If the converted string is an empty string, *cbufsize_ret gets set to 0, but *cbufptr_ret still gets initialized (to a dummy malloced buffer).

A non-zero nullterminate places a trailing \0 character after the converted string (this is included in *cbufsize_ret).

Converting between character sets and unicode

unicode_convert_tou_init() converts character text into a char32_t buffer. It works just like unicode_convert_tocbuf_init(), except that only the source character set gets specified and the output buffer is a char32_t buffer. nullterminate terminates the converted unicode characters with a U+0000.

unicode_convert_fromu_init() converts char32_ts to the output character set, and also works like unicode_convert_tocbuf_init(). Additionally, in this case, unicode_convert_uc() works just like unicode_convert() except that the input sequence is a char32_t sequence, and the count parameter is th enumber of unicode characters.

One-shot conversions

unicode_convert_toutf8() converts the specified text in the specified text into a UTF-8 string, returning a malloced buffer. If error is not NULL, even if unicode_convert_toutf8() returns a non NULL value *error gets set to a non-zero value if a character conversion error has occurred, and some characters could not be converted.

unicode_convert_fromutf8() does a similar conversion from UTF-8 text to the specified character set.

unicode_convert_tobuf() does a similar conversion between two different character sets.

unicode_convert_tou_tobuf() calls unicode_convert_tou_init(), feeds the character string through unicode_convert(), then calls unicode_convert_deinit(). If this function returns 0, *uc and *ucsize are set to a malloced buffer+size holding the unicode char array.

unicode_convert_fromu_tobuf() calls unicode_convert_fromu_init(), feeds the unicode array through unicode_convert_uc(), then calls unicode_convert_deinit(). If this function returns 0, *c and *csize are set to a malloced buffer+size holding the char array.